Fix HDMI audio detection on Pi4 with dual HDMI ports

[Alex1981-tech]

Summary

• Pi4 has two HDMI ports (HDMI-A-1 / HDMI-A-2), but VLC audio always targets default:CARD=vc4hdmi0 regardless of which port has a display connected. This causes no audio when using HDMI-A-2.
• The default:CARD= ALSA device name can be unreliable when PulseAudio or dmix is involved.

Changes

• Add _detect_hdmi_audio_device() function that reads /sys/class/drm/cardN-HDMI-A-N/status to detect which HDMI port is connected
• Switch from default:CARD= to sysdefault:CARD= for direct hardware access
• Apply auto-detection for Pi4 and Pi5 HDMI output
• Headphones and Pi1-3 paths are unchanged

Test plan

• Tested on Pi4 Model B Rev 1.5, Raspberry Pi OS, single HDMI connected to HDMI-A-1
• Audio plays correctly through HDMI
• Fallback works when no HDMI status file is found
• Needs testing with HDMI-A-2 only
• Needs testing on Pi5

[nicomiguelino]

@Alex1981-tech, thank you for opening a pull request! I will test the changes when I get the chance.
In the meantime, please continue testing your changes and fix failures in checks—linting & formatting, tests.

[Alex1981-tech]

CI update: Fixed ruff format. The remaining test failure (test_add_asset_url) is a pre-existing flaky Selenium integration test unrelated to this PR — it fails on other PRs as well (e.g. Dependabot).

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[vpetersson]

Hi @Alex1981-tech — thanks again for this fix! Closing this PR in favor of #2811, which supersedes it.

Master moved on since you opened #2667 (the media_player.py refactor turned get_alsa_audio_device() into a top-level helper shared between the MPV and VLC players), so the branch became conflicted. I've rebased your work onto current master, resolved the conflict against the new structure, and added test coverage for _detect_hdmi_audio_device(). Your commit and authorship are preserved on the new PR.

Closing this PR — please follow along on #2811.